This paper describes how a soft contact lens can be used to harvest cells from the surface of the corneal epithelium. The procedure is called contact lens cytology (CLC). Cells were removed from a soft contact lens by irrigation and stained with acridine orange. Two methods for the measurement of cell size are described. First, cell size was measured using a computer-assisted technique, which calculated the area of the cell from its outline. The second method was simpler in that it required only a single measurement of the longest dimension of the cell (the cell length). To test the validity of this simpler method, cell area was compared with cell length in 185 cells. The resulting correlation (r = 0.92) suggests that the size of shed cells can be described adequately using cell length in place of the more time-consuming measurement of cell area. A mathematical relation can be used to convert cell length to cell area so that results from experimenters using different measures of size can be compared. When a large pool of cells collected by CLC was divided into four aliquots and cell length measured by two observers on two different days, there were no significant differences between observers or days. Thus, the technique does not depend on one observer, and it is unaffected by a 24-h delay in measurement. Cells were harvested from the corneal epithelium of normal human subjects. The number of cells collected from any single removal of the contact lens had a range of 10 to 175 cells, and a mean of 66.8 +/- 40.4 (N = 46). Cell length was measured and plotted as frequency histograms for both eyes of each subject. The range in cell length was from 10 to 80 microns. The mean cell length for individual subjects had a low of 26.5 +/- 9.0 microns and a high of 44.2 +/- 10.2 microns, with a grand mean for all right eyes of 36.0 +/- 5.1 microns, and a grand mean for all left eyes of 34.6 +/- 5.2 microns. The mean for all eyes was 35.3 +/- 5.1 microns. Composite histograms were created with the combined data from the 23 right eyes (N = 1310 cells), and the 23 left eyes (N = 1765 cells). Individual histograms and the composite histograms were not normally distributed. Peaks in the distributions suggest the presence of different subpopulations of cells, lending support to the hypothesis that there is more than one mechanism for cell shedding.